Capybara: Ambiguous match, found 2 elements matching xpath "/html"

Created on 21 Nov 2012 · 22Comments · Source: teamcapybara/capybara

I'm using capybara (2.0.1) and I'm getting this issue whenever I write an expectation such as:

page.has_content?

However, everything works if I rewrite this to:

find('body').has_content?

I'm not sure if this is the intended functionality or if I've misunderstood the new changes.

Source

tonydaly

👍2

Most helpful comment

Done some more debugging on the issue. Did a sanity check and there aren't multiple HTML tags on the page, but it seems like having a doctype of <!DOCTYPE html> is causing the issue. After removing the doctype everything passes.

tonydaly on 21 Nov 2012

👍3

All 22 comments

Sounds like you have multiple HTML tags on your page. As far as I know that's not allowed by the HTML spec. See: http://dev.w3.org/html5/spec/single-page.html#the-html-element

jnicklas on 21 Nov 2012

🎉1

tonydaly on 21 Nov 2012

👍3

@tonydaly that should not affect anything.

Nokogiri::HTML("<!DOCTYPE html><html></html").xpath("/html").size # => 1

jnicklas on 21 Nov 2012

@tonydaly, maybe your closing </html> tag is missing the slash? Though experimenting with Nokogiri doesn't appear to yield those kinds of results either, what driver are you using?

jnicklas on 21 Nov 2012

Erugh ignore this. There was a tag outside of the closing </html>.

tonydaly on 27 Nov 2012

🎉1 👍1

@tonydaly thanks for following up. Had the same thing because another dev put extra stuff in there and it would have been the last thing I checked (Literally, it's at the end of the file!). :clap:

jakeonrails on 3 May 2013

For me it's this one causing the trouble:

<!DOCTYPE html>
<!--[if lt IE 7]><html class="lt-ie7"> <![endif]-->

<html>
...
</html>

As you can see even this parser recognizes that the multiple <html> tags are just comments and Capybara maybe shouldn't raise an Exception :) I'm using 2.1.0

mmzoo on 18 Jul 2013

👍1

Ah, my bad. There IS another between all those comments I have :)

mmzoo on 18 Jul 2013

@tonydaly nice catch. I added some JavaScript outside of the trailing HTML tag (in HAML), and it caused the same failure. Thanks for the conversation/fix :+1:

calebhaye on 8 Mar 2014

👍2

I had a similar problem as @calebhaye, script tag outside of the body tag.

taylorkearns on 30 May 2014

i was using haml....i missed indention right after %hmtl that caused it.

amba178 on 21 Aug 2014

I have the same problem because I am using Bullet gem. Bullet gem appends content outside the html make the html malformed.

channainfo on 24 Oct 2014

@tonydaly (y) my javascript code was outside the html tag. (As slim works with indentation, so after indenting my code inside html it is fine)

taimoor on 17 Dec 2014

Just had this issue and I too had a

Just in case someone hits this problem while debugging a Rails App: The Rails web console was causing it in my case. When I took console out of my controller, the tests went green again.

TStrothjohann on 2 Apr 2015

@TStrothjohann thanks buddy, you save my day.

philidor-green on 11 Aug 2015

@TStrothjohann That was my problem too, upgrading web-console fixes this. See this web-console issue.

mooman on 23 Feb 2016

I had the same problem. Thanks for sharing the solution.

ElizabethRamos on 2 Jul 2017

Thanks so much for the discussion here. It helped point me in the right direction. I found that my application.html.slim file was improperly nested. That is there were some elements outside of the html open and closing tags.

jr180180 on 1 Jun 2018

For me it's this one causing the trouble:
<!DOCTYPE html>


<html>
...
</html>
As you can see even this parser recognizes that the multiple <html> tags are just comments and Capybara maybe shouldn't raise an Exception :) I'm using 2.1.0

https://html.com/tags/doctype/#HTML5_Documents

chabgood on 4 Mar 2019

@chabgood I have no clue why you're adding a link about doctype to a comment from 6 years ago - I'm locking this issue. If you have an actual issue please create a new one with enough details to reproduce it.